CN108460727A - A kind of image split-joint method based on perspective geometry and SIFT feature - Google Patents
A kind of image split-joint method based on perspective geometry and SIFT feature Download PDFInfo
- Publication number
- CN108460727A CN108460727A CN201810262297.7A CN201810262297A CN108460727A CN 108460727 A CN108460727 A CN 108460727A CN 201810262297 A CN201810262297 A CN 201810262297A CN 108460727 A CN108460727 A CN 108460727A
- Authority
- CN
- China
- Prior art keywords
- image
- matching
- point
- feature
- characteristic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 239000011159 matrix material Substances 0.000 claims abstract description 29
- 230000009466 transformation Effects 0.000 claims abstract description 25
- 238000000746 purification Methods 0.000 claims abstract description 11
- 238000005498 polishing Methods 0.000 claims abstract description 7
- 238000000605 extraction Methods 0.000 claims abstract description 6
- 238000003384 imaging method Methods 0.000 claims description 4
- 230000001502 supplementing effect Effects 0.000 claims 1
- 238000010845 search algorithm Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of image split-joint method based on perspective geometry Yu SIFT feature matching double points, this method shoots two width first has the image of overlapping region, stitching image is treated to carry out SIFT feature extraction and carry out Feature Points Matching with K D tree search algorithms, RANSAC algorithms are used to carry out characteristic point purification to reject the matching double points to make mistake again, transformation matrix is calculated if the characteristic matching point after purification is to being more than 8 pairs, the projection matching point for extracting respective numbers according to the overlapping region known to two images if the characteristic matching point after purification is to less than 8 pairs calculates transformation matrix completion image registration to polishing 8 to matching double points, image co-registration is carried out using multiresolution algorithm to the image after registration, finally export stitching image.Using method proposed by the present invention carry out image mosaic can solve because characteristic matching point is to less so that image registration is failed the case where, while image mosaic works well.
Description
Technical field
The invention belongs to technical field of image processing, and in particular to a method of being used for Panorama Mosaic.
Background technology
With the development of computer technology, Panorama Mosaic technology has obtained extensive research and development.In panorama sketch
As the step of in splicing, image registration is most critical in image mosaic, the success or failure of image mosaic are directly influenced.Image is matched
Quasi- method includes mainly relevant based on phase, based on geometric areas and feature based merging algorithm for images.Based on phase
Input image sequence is first carried out Fourier transform by the relevant merging algorithm for images in position first, is then utilized mutual after image transformation
Phase information in power spectrum calculates the relative displacement between image to carry out image registration.Based on the relevant image of geometric areas
Stitching algorithm be by image slices vegetarian refreshments gray level, to the partial geometry subregion of input picture carry out related operation come into
Row image registration.The merging algorithm for images of feature based extracts the feature of image to be spliced first, then by characteristic matching come
Complete image registration.The image registration techniques of wherein feature based are a hot spot of image procossing research field in recent years, base
In the image split-joint method of feature be the most common method in image mosaic field.The method for registering of feature based needs first calculating figure
Accurate transformation matrix as between, the transformation matrix for how obtaining the position of image registration or calculating between image is that image is matched
Accurate key.In Panorama Mosaic technology, most classical method is David Lowe propositions based on scale invariant feature
The method for converting (SIFT).For this method by being matched to image zooming-out SIFT feature, SIFT features are flat to image
Shifting, rotation, scaling, brightness change all have invariance, while also having good Shandong to visual angle change, affine transformation, noise
Stick, the practicality is very strong, is the most common algorithm in image mosaic field of feature based.But work as characteristics of image unobvious
When, such as sky, ocean, meadow scenery image, the feature that this method can be extracted is few, for feature calculation effect it is unknown
It is aobvious, even it can not complete image mosaic sometimes.There is presently no which kind of algorithms can all obtain good under any circumstance
With effect, therefore which kind of image split-joint method is used, depends on the practical ranges of specific algorithm and the content of image.
Invention content
Present invention aim to solve when characteristics of image unobvious, because extractible feature is few, not completing
The problem of image mosaic, provides a kind of image mosaic side based on perspective geometry Yu SIFT feature matching double points for image mosaic
Method.
The present invention specific implementation step be：
Step 1：Fixed camera position, which is continuously shot two width, has the image of overlapping region, and overlapping region is made to account for image surface
30% to 50% long-pending and clear overlapping region position.
Step 2：The characteristic point of two images to be spliced is extracted, and carries out Feature Points Matching, the feature after matching is clicked through
Row purification is to reject the characteristic matching to make mistake point pair.
Step 3：Transformation matrix is calculated using image projection transformation model.Whether judging characteristic matching double points number is more than 8
It is right, image transformation matrix is then calculated more than 8 pairs, is then projected according to overlapping region position acquisition known to two images less than 8 Duis
Matching double points, it is right to polishing 8 to randomly select appropriate projection matching point, calculates image transformation matrix and completes image registration.
Step 4：Image co-registration is carried out to the image after registration using multiresolution algorithm, finally exports stitching image.
In step 3 projection matching point to be corresponding same outdoor scene imaging point under different camera sites two width difference
Pixel on image, acquisition methods are as follows：
According to known two images overlapping region position, the overlapping region in piece image is chosen in two images
Vertex obtains the edge line of overlapping region, using the midpoint of the overlapping region edge line segment as subpoint, with adjacent figure to be spliced
The subpoint of corresponding position as in constitutes projection matching point pair.
Compared with prior art, the beneficial effects of the invention are as follows：
For the deficiency of SIFT feature extraction algorithm, when image feature information unobvious, parts of images is because of extractible spy
Sign point is less, the shortcomings that can not calculating transformation matrix, and set forth herein a kind of based on perspective geometry and SIFT feature matching double points
Image split-joint method, this method calculate transformation matrix using projection matching point, can greatly improve the success rate of image mosaic, together
When image mosaic work well.
Specific implementation mode
Below in conjunction with the specific implementation mode of the description of the drawings present invention, it should be understood that the implementation for showing and describing in attached drawing
Mode is merely exemplary, it is intended that is illustrated the principle of the present invention and method, and is not intended to limit the scope of the invention.
As shown in Figure 1, a kind of image mosaic side based on perspective geometry Yu SIFT feature matching double points proposed by the present invention
Method, specific implementation step are：Step 1：Fixed camera position, which is continuously shot two width, has the image of overlapping region, makes overlay region
Domain accounts for the 30% to 50% of image area and clear overlapping region position.
Step 2：The characteristic point of two images to be spliced is extracted, and carries out Feature Points Matching, the feature after matching is clicked through
Row purification is to reject the characteristic matching to make mistake point pair.
Step 3：Transformation matrix is calculated using image projection transformation model.Whether judging characteristic matching double points number is more than 8
It is right, image transformation matrix is then calculated more than 8 pairs, is then projected according to overlapping region position acquisition known to two images less than 8 Duis
Matching double points, it is right to polishing 8 to randomly select appropriate projection matching point, calculates image transformation matrix and completes image registration.
Step 4：Image co-registration is carried out to the image after registration using multiresolution algorithm, finally exports stitching image.
Projection matching point is to obtaining schematic diagram as shown in Fig. 2, dash area is the overlapping of two images in figure in step 3
Region.A1A3A6A8 and B1 B3B6B8 are the vertex of shadow region, and A2A4A5A7 and B2B4B5B7 is in the line segment of shadow region
Point.According to image imaging geometry principle, figure midpoint A1 and point B1 be same realistic picture picture point under different shooting angles in two width
Imaging point on different images, according to perspective geometry principle, A1 and B1 is one-to-one relationship, is a pair of of projection matching point, together
Reason, A2 and B2, A3 and B3, A4 and B4, A5 and B5, A6 and B6, A7 and B7, A8 and B8 are projection matching points pair.
Characteristic matching point in step 3 after purification is to less than 8 clock synchronizations, using projection matching point to polishing, such as Fig. 3 institutes
Showing, point A4B4, A5B5, A6B6 are using the characteristic matching point pair of SIFT feature algorithm extraction, remaining is projection matching point pair,
Projection matching point pair is with characteristic matching point to constituting 8 pairs of matching double points together.Image is calculated using image projection transformation model to become
Change matrix.
Image projection transformation model calculates as follows：
If image A to be spliced is with a pair of of corresponding points on BWithThey are full
Sufficient epipolar geometry constraints, epipolar geometry constraints are described using F matrix：
Wherein,
] when there are n to corresponding points by image A to be spliced and B, matrix A is constructed,
Make Af=0
F=[F11 F12 F13 F21 F22 F23 F31 F32 F33]T
Analysis is carried out to above formula to find, it, can be in the hope of matrix f as the logarithm n >=8 of corresponding points.Therefore work as known 8 groups
With characteristic point clock synchronization, so that it may with linear solution f.In order to solve this over-determined systems, need to carry out SVD decomposition to matrix A, i.e.,
A=UDVT, and f is equal to the feature vector corresponding to the minimum singular value of A.Use the basis matrix F that f is built can't be as
Final result will also ensure that the basis matrix acquired is singular matrix, because only that unusual basis matrix can just make polar curve phase
It meets at a bit.The constraint for being 2 into row rank to matrix F, has
F=Udiag (s1 s2 s3)VT
Work as s3When=0, there is the estimation of matrix F
The validity of institute's extracting method of the present invention is verified below by specific embodiment.It is pointed out that the embodiment
It is only exemplary, is not intended to limit the scope of application of the present invention.
Fixed camera translates angle and shoots image of two width with 30% overlapping region, and picture material is dull, makes shooting
Image feature information is less.As shown in Figure 4.
Feature extraction first is carried out with SIFT algorithms to this two images and goes out characteristic point, then is carried out using K-D tree search algorithms
Feature Points Matching rejects the match point to make mistake to the characteristic point after matching to carrying out characteristic point purification using RANSAC algorithms.
Since the characteristic matching point after purification is to right less than 8, the projection matching point polishing of respective numbers is randomly selected, is calculated
Image transformation matrix completes image registration.
Image after registration merges image using multi-resolution Fusion technology, exports stitching image.Image is spelled
It is as shown in Figure 5 to connect effect.
Description of the drawings
Fig. 1 is a kind of stream based on perspective geometry Yu the image split-joint method of SIFT feature matching double points proposed by the present invention
Cheng Tu
Fig. 2 is two image projection geometric match figures to be spliced
Fig. 3 is two image projection geometry to be spliced and SIFT feature matching figure
Fig. 4 is the two images acquired using image pickup method of the present invention
Fig. 5 is the design sketch to the two images splicing of acquisition using image split-joint method of the present invention.
Claims (4)
1. a kind of image split-joint method based on perspective geometry Yu SIFT feature matching double points, which is characterized in that this method includes
Following steps：
Step 1：Fixed camera position, which is continuously shot two width, has the image of overlapping region so that overlapping region accounts for single image
30% to 50% and clear overlapping region position of area.
Step 2：Feature point extraction is carried out to two images to be spliced, and carries out Feature Points Matching, the feature after matching is clicked through
Row purification is to reject the characteristic matching to make mistake point pair.
Step 3：Transformation matrix is calculated using image projection transformation model, counts whether effective characteristic matching point surpasses number
8 pairs are crossed, image transformation matrix is then calculated more than 8 pairs, is then thrown according to overlapping region position acquisition known to two images less than 8 Duis
Shadow matching double points, it is right to polishing 8 using projection matching point, with projection matching point to supplementing, make matching characteristic point to being not less than 8
It is right, it calculates image transformation matrix and completes image registration.
Step 4：The image after registration is merged using multiresolution algorithm, exports stitching image.
2. the image split-joint method according to claim 1 based on perspective geometry Yu SIFT feature matching double points, feature
It is, the characteristic point of two images to be spliced is extracted in the step 2 using SIFT feature extraction algorithm, using multi-C vector
Nearest neighbor search method K-D tree algorithms carry out Feature Points Matching, and feature is carried out using RANSAC algorithms to the characteristic point after matching
Point purification, to reject the characteristic matching point pair to make mistake.
3. the image split-joint method according to claim 1 based on perspective geometry Yu SIFT feature matching double points, feature
It is, projection matching point is to being two width difference figures that the corresponding imaging point of same outdoor scene is shot in different angle in the step 3
Picture point as in, acquisition methods are as follows：
According to known two images overlapping region position, the vertex of overlapping region, the midpoint conduct of overlapping edge line segment are chosen
The subpoint of corresponding position in subpoint, with adjacent image to be spliced constitutes projection matching point pair.
4. the image split-joint method according to claim 1 based on perspective geometry Yu SIFT feature matching double points, feature
It is, the characteristic matching point in the step 3 after purification less than 8 clock synchronizations to that can not calculate image transformation matrix, using suitable
It measures projection matching point and image registration is completed to calculating image transformation matrix to polishing 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810262297.7A CN108460727A (en) | 2018-03-28 | 2018-03-28 | A kind of image split-joint method based on perspective geometry and SIFT feature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810262297.7A CN108460727A (en) | 2018-03-28 | 2018-03-28 | A kind of image split-joint method based on perspective geometry and SIFT feature |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108460727A true CN108460727A (en) | 2018-08-28 |
Family
ID=63237104
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810262297.7A Pending CN108460727A (en) | 2018-03-28 | 2018-03-28 | A kind of image split-joint method based on perspective geometry and SIFT feature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108460727A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886878A (en) * | 2019-03-20 | 2019-06-14 | 中南大学 | A kind of infrared image joining method based on by being slightly registrated to essence |
| CN110232656A (en) * | 2019-06-13 | 2019-09-13 | 上海倍肯机电科技有限公司 | A kind of insufficient image mosaic optimization method of solution characteristic point |
| CN110852988A (en) * | 2019-09-27 | 2020-02-28 | 广东电网有限责任公司清远供电局 | Method, device and equipment for detecting self-explosion of insulator string and storage medium |
| CN110852986A (en) * | 2019-09-24 | 2020-02-28 | 广东电网有限责任公司清远供电局 | Method, device and equipment for detecting self-explosion of double-string insulator and storage medium |
| CN111553870A (en) * | 2020-07-13 | 2020-08-18 | 成都中轨轨道设备有限公司 | Image processing method based on distributed system |
| CN112258395A (en) * | 2020-11-12 | 2021-01-22 | 珠海大横琴科技发展有限公司 | Image splicing method and device shot by unmanned aerial vehicle |
| CN114220068A (en) * | 2021-11-08 | 2022-03-22 | 珠海优特电力科技股份有限公司 | Method, device, equipment, medium and product for determining on-off state of disconnecting link |
| CN116109852A (en) * | 2023-04-13 | 2023-05-12 | 安徽大学 | Quick and high-precision feature matching error elimination method |
-
2018
- 2018-03-28 CN CN201810262297.7A patent/CN108460727A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886878A (en) * | 2019-03-20 | 2019-06-14 | 中南大学 | A kind of infrared image joining method based on by being slightly registrated to essence |
| CN110232656A (en) * | 2019-06-13 | 2019-09-13 | 上海倍肯机电科技有限公司 | A kind of insufficient image mosaic optimization method of solution characteristic point |
| CN110232656B (en) * | 2019-06-13 | 2023-03-28 | 上海倍肯智能科技有限公司 | Image splicing optimization method for solving problem of insufficient feature points |
| CN110852986A (en) * | 2019-09-24 | 2020-02-28 | 广东电网有限责任公司清远供电局 | Method, device and equipment for detecting self-explosion of double-string insulator and storage medium |
| CN110852988A (en) * | 2019-09-27 | 2020-02-28 | 广东电网有限责任公司清远供电局 | Method, device and equipment for detecting self-explosion of insulator string and storage medium |
| CN111553870A (en) * | 2020-07-13 | 2020-08-18 | 成都中轨轨道设备有限公司 | Image processing method based on distributed system |
| CN112258395A (en) * | 2020-11-12 | 2021-01-22 | 珠海大横琴科技发展有限公司 | Image splicing method and device shot by unmanned aerial vehicle |
| CN114220068A (en) * | 2021-11-08 | 2022-03-22 | 珠海优特电力科技股份有限公司 | Method, device, equipment, medium and product for determining on-off state of disconnecting link |
| CN114220068B (en) * | 2021-11-08 | 2023-09-01 | 珠海优特电力科技股份有限公司 | Method, device, equipment, medium and product for determining disconnecting link switching state |
| CN116109852A (en) * | 2023-04-13 | 2023-05-12 | 安徽大学 | Quick and high-precision feature matching error elimination method |
| CN116109852B (en) * | 2023-04-13 | 2023-06-20 | 安徽大学 | Quick and high-precision image feature matching error elimination method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108460727A (en) | A kind of image split-joint method based on perspective geometry and SIFT feature | |
| Pritchett et al. | Wide baseline stereo matching | |
| CN104200461B (en) | The remote sensing image registration method of block and sift features is selected based on mutual information image | |
| CN106919944A (en) | A kind of wide-angle image method for quickly identifying based on ORB algorithms | |
| Wu | Creating panoramic images using ORB feature detection and RANSAC-based image alignment | |
| Mistry et al. | Image stitching using Harris feature detection | |
| Lin et al. | Aligning images in the wild | |
| CN101488224B (en) | Characteristic point matching method based on relativity measurement | |
| Hua et al. | Image stitch algorithm based on SIFT and MVSC | |
| CN103258321A (en) | Image stitching method | |
| Ying et al. | Self-calibration of catadioptric camera with two planar mirrors from silhouettes | |
| Wang et al. | Automatic fundus images mosaic based on SIFT feature | |
| CN110120013A (en) | A kind of cloud method and device | |
| Hong et al. | Panoramic image reflection removal | |
| CN114066954B (en) | Feature extraction and registration method for multi-modal image | |
| Pedone et al. | Blur invariant translational image registration for $ N $-fold symmetric blurs | |
| CN105654479A (en) | Multispectral image registering method and multispectral image registering device | |
| Cho et al. | Automatic Image Mosaic System Using Image Feature Detection and Taylor Series. | |
| Chuang et al. | Rectified feature matching for spherical panoramic images | |
| CN111047513A (en) | Robust image alignment method and device for cylindrical panoramic stitching | |
| Belongie et al. | Structure from periodic motion | |
| Jiao et al. | Highlight removal for camera captured documents based on image stitching | |
| Tang et al. | Sequential far infrared image mosaic using coarse-to-fine scheme | |
| Lourakis | Plane metric rectification from a single view of multiple coplanar circles | |
| Bunnun et al. | Integrating 3D object detection, modelling and tracking on a mobile phone |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180828 |